Recorded activity, sustained adoption and classified Craft: v7.5_

Activity/60

The unchanged v7.3 activity composite earns 60% of its value, up to 60 points. Each of Delivery, Quality, Consistency and Breadth contributes up to 15 points. Activity uses the declared window of 365 UTC dates.

The badge radar shows credited score points: Delivery, Quality, Consistency and Breadth contribute up to 15 each; Adoption contributes up to 40. All axes use the same points scale, and the outline shows these unequal limits. Exact contributions add up to the headline before it is rounded down. Craft is reported separately and contributes no headline points. The detailed activity diagnostics remain 0–100. A test artifact does not prove a passing test. Archetype describes activity only.

Outcomes/40

Optionally connect PostHog or, from v7.6, Google Analytics with read-only access. One product uses the minimum of exactly six completed calendar-month counts of its active users: PostHog counts active analytics identifiers in UTC months, and Google Analytics counts active users in the property reporting time zone. The current partial month is excluded; the six-month period must span at least 180 days within the activity window.

Magnitude interpolates linearly from 0 to 100 identifiers, then linearly in logarithmic count between adjacent positive anchors. Counts above 10 million are capped at 40 points.

Credited outcome points equal the lesser of magnitude and verified project participation. In the last 180 UTC days, 12 distinct project/acceptance-date units across 6 active ISO weeks permit 10 points; 24 units across 12 weeks permit 30. The 40-point cap also requires 60 units across 26 weeks in the 365-date window. Incomplete attribution or source collection cannot qualify.

12/6 → 10 · 24/12 → 30 · 24/12 + 60/26 → 40

Subject-authored accepted work must match the linked repository and a successful production deployment. The domain association is provider-reported and ancestry-checked; it does not measure contribution complexity or causal share.

Six completed UTC months, one product. Analytics identifiers are not people. Query completion does not prove ingestion completeness; deployment association does not prove causal impact.

Evidence band

Bands describe declared evidence criteria. An activity-only score can reach 60 and remains Established. Adoption and higher bands require positive credited outcomes. Exceptional requires demanding recorded adoption criteria; it is not a worldwide percentile.

• Limited: <30
• Established: 30–<60; O=0: ≥30
• Adoption: 60–<80; O>0
• Substantial: 80–<95; O>0
• Exceptional: 95–100; O>0

The headline is the floor of the exact total. Components are not floored before addition. Each sealed receipt carries its exact total, displayed integer and evidence band.

Availability and recovery

An unmeasured outcome earns zero accounting points and remains explicitly unmeasured. A valid returned zero is a measured zero. Product observation can be measured while contribution eligibility is unmeasured. Other outcomes, including revenue and noninstrumented reach, are unsupported by this adapter.

Computation must be less than 24 hours old at publication. Expiry is last refresh plus 24 hours; capture does not extend it. Expiry and a new UTC reference month mark the observation stale. Configuration changes block old-generation credit; missing new-policy evidence shows collecting or unavailable instead of an old score.

How your archetype is assigned

Every profile receives an archetype from the four exact recorded activity dimensions. Open evidence ranges do not block assignment. With no recorded activity, the profile starts as Emerging. Craft and product adoption never change the activity archetype. Historical signed receipts keep their original values.

Craft

Craft is a separate report estimate and never changes the activity or outcome points.

Craft estimates outcomes among classified sessions: 100 × (fully achieved + 0.7 × mostly achieved + 0.3 × partially achieved) / classified sessions, rounded to the nearest integer. Explicit failures count in the denominator with zero credit. Unknown and unclassified sessions are excluded. For 39 fully achieved, 39 mostly achieved, 8 partially achieved and 1 failed session, 100 × 68.7 / 87 = 78.96551724137932, displayed as 79. With 1 unknown and 746 unclassified sessions, coverage is 87 of 834 sessions classified (10.43%). This score does not imply success in the omitted sessions or measure general ability. There is no minimum coverage penalty: zero classified sessions means insufficient data without a score; classified failures alone can give a measured zero. Craft has zero core weight.

Receipts and historical policies

Every consumer copies the same sealed total, band, policy and revision. v7.5 records the classified-session calculation basis, sample size, coverage and report period. Its unchanged activity and adoption rules match v7.4. Historical v7, v7.2, v7.3 and v7.4 receipts keep their original arithmetic: the same example remains 8 in a v7.4 receipt. New v7.5 issuance derives 79 from retained valid report aggregates without reupload or rewriting the old report or receipt. Public aggregates enable arithmetic replay; signatures prove recorded issuance, not source accuracy. Policies and windows are not directly comparable.